Socio-Technical Landscapes of Artificial Intelligence in India: A Comprehensive Analysis of Public Perception, Ethical Governance, and Market Maturity (2025-2026)

Socio-Technical Landscapes of Artificial Intelligence in India: A Comprehensive Analysis of Public Perception, Ethical Governance, and Market Maturity (2025-2026)

Research Dissertation submitted for DPGU CONFERENCE Date: February 21, 2026.

Akash Dnyaneshwar Wagawad

Department of Computer Science

Dr. D. Y. Patil Arts, Commerce and Science College- 411018 Pune, India

Durgeshwari Sanjay Rachote

Department of Computer Science

Dr. D. Y. Patil Arts, Commerce and Science College- 411018 Pune, India

Abstract – The period of 20252026 marks a definitive "sovereign turn" in Indias engagement with Artificial Intelligence (AI). Moving beyond the passive adoption of Western foundational models, India has architected a unique "Techno-Legal" ecosystem characterized by state-led Digital Public Infrastructure (DPI) and aggressive private-sector application. This dissertation presents a mixed-methods study of this transition, utilizing data from 52 institutional reports, legislative texts (specifically the IT Rules Amendment 2026), and a stratified survey of public sentiment.

The central thesis of this research is that India is constructing a "Third Way" of AI governance-distinct from the laissez-faire American model and the rights-centric European model. This "Indian Model" is defined by Pragmatic Dualism: simultaneously fostering hyper-growth in the B2B agentic AI sector (projected $130B by 2032) while enforcing draconian, post-facto accountability measures for social harms (e.g., the 3- hour deepfake takedown mandate). However, this rapid maturity masks a critical "Socio-Technical Gap": a disconnect between high enterprise adoption (87%) and low public algorithmic literacy (14%). This paper argues that without immediate sociological interventions, this gap risks calcifying into a new form of digital exclusion, where the benefits of AI are privatized by the "cognitive elite" while the risks of surveillance and displacement are socialized among the masses.

Keywords: Socio-Technical Systems, Digital Public Infrastructure (DPI), Agentic AI, Algorithmic Governance, IT Rules 2026, Deepfakes, Sovereign Compute.

1: INTRODUCTION

1. Research Context: The Global and Local AI Narrative

   By early 2026, the global discourse on Artificial Intelligence has shifted from "capability" to "deployment." The era of "shock and awe" regarding Generative AIs abilities (2023 2024) has ceded ground to the pragmatic integration of "Agentic AI"-systems that do not merely generate text or images but execute complex, multi-step workflows with autonomy.

   In this global theater, India occupies a paradoxical position.

   The Stanford Global AI Vibrancy Tool (2025) ranks India third globally, trailing only the United States and China [1]. Yet, this ranking is heavily skewed by "human capital" and "diffusion" metrics, while "research" and "infrastructure" lag.

   The narrative of AI in India is thus one of asymmetry: a surplus of talent and application developers juxtaposed against a deficit of indigenous foundational models and high-end compute (GPUs).

   However, the states intervention in late 2025, via the operationalization of the IndiaAI Mission (10,370 crore), has begun to correct this imbalance.

   By subsidizing compute at 65/hour for startups, the Indian state has signaled that AI is not merely a market commodity but a "Public Good"-akin to highways or electricity [2]. This dissertation explores how this state-led developmentalist approach interacts with the chaotic, hyper-capitalist energy of India's private tech sector and the complex, pluralistic fabric of Indian society.

2. Statement of the Problem: The "Trust-Knowledge Paradox"

   The core problem addressing this research is the "Trust- Knowledge Paradox" (TKP). Preliminary data suggests that while Indian citizens display some of the highest levels of "trust" in AI systems globally (KPMG, 2025), this trust is not correlated with "knowledge" or "understanding" of how these systems function [3].

   1. High Trust/Adoption: 87% of Indian enterprises have adopted AI, and 66% of citizens use AI-integrated

      tools daily (often unknowingly via UPI or platforms like Zomato).

   2. Low Literacy: Only 14% of the population can accurately define "algorithm" or explain how their data is used to train these models.

   3. The Risk: This "blind trust" creates a vulnerability to Algorithmic Harm. If citizens trust AI implicitly because they associate it with state success (e.g., UPI success), they are less likely to question biased loan denials, facial recognition errors, or synthetic propaganda.

   4. The problem is further compounded by the IT Intermediary Rules Amendment (2026). While the law mandates strict takedowns of deepfakes, the enforcement relies on user reporting. If users lack the visual literacy to identify a deepfake, the legal safeguard becomes moot. Thus, the socio-technical landscape is currently fragile: built on the shaky foundation of uncritical acceptance rather than informed consent.

3. Research Questions (RQs)

   To address this paradox, this dissertation poses four primary research questions:

   • RQ1 (Market Dynamics): How has the shift from "Generative AI" to "Agentic AI" in 20252026 altered the economic structure of Indias IT and BPM sectors?

   • RQ2 (Governance): To what extent does the "Techno- Legal" framework of 2026 (specifically the 3-hour takedown rule) effectively mitigate the risks of Synthetically Generated Information (SGI) compared to global standards?

   • RQ3 (Public Perception): What is the correlation between "Digital Literacy" and "Algorithmic Anxiety" among diverse Indian demographics (urban vs. rural, Gen Z vs. Boomers)?

   • RQ4 (Sovereignty): Can the "IndiaAI Mission" and initiatives like BharatGen successfully create "Compute Sovereignty," or will India remain dependent on Western semiconductor supply chains?

4. Significance of the Study

• For Policy Makers: This study provides a critical audit of the IndiaAI Mission and the IT Rules 2026,

  offering data-driven recommendations for the upcoming Digital India Act (2027).

• For Industry: By analyzing the "Agentic Shift," this

  research offers a roadmap for GCCs and startups to navigate the talent crunch and the ethical compliance landscape.

• For Academia: It contributes to the field of Post- Colonial Science and Technology Studies (STS),

  challenging the "Universalist" view of AI ethics by highlighting the specific caste, linguistic, and regional nuances of the Indian experience.

  2: LITERATURE REVIEW & THEORETICAL FRAMEWORK

  1. Theoretical Framework: Socio-Technical Systems (STS)

     This research is grounded in the Socio-Technical Systems (STS) theory, originally proposed by Trist and Bamforth (1951) and adapted for the digital age by scholars like Kate Crawford and Langdon Winner.

     1. The "Social Shaping of Technology" in India

        STS posits that technology is not an isolated artifact but an assemblage of social, political, and economic forces. In the Indian context, AI cannot be understood merely s code (Python/PyTorch). It must be understood as:

        • Political: A tool for state capacity building (e.g., using AI for tax evasion detection).

        • Cultural: A site of linguistic struggle (English vs. Indic languages).

        • Economic: A mechanism for labor arbitrage (moving from low-cost coding to high-value AI reasoning).

        We reject "Technological Determinism" (the idea that AI inevitably leads to specific outcomes). Instead, we adopt a "Mutual Shaping" perspective: Indian society shapes AI (through unique data patterns, regulatory demands, and cultural usage), just as AI shapes Indian society (through labor displacement and surveillance).

     2. The "Algorithmic Colonization" Hypothesis

  A subset of the literature review focuses on the concept of "Algorithmic Colonization" (Mohamed et al., 2020) [4]. This theory suggests that importing Western Foundation Models (like GPT-4 or Claude) imposes Western values, biases, and "ground truths" onto the Global South.

  Example: An LLM trained on Western internet data might identify a "wedding" only with white dresses and churches, flagging Hindu or Islamic symbols as "outliers" or "noise."

  Relevance: This framework is crucial for analyzing the BharatGen initiative. Is BharatGen merely a copy of Western models, or does it represent a genuine "de- colonial" epistemic break?

  1. From Co-pilots to Agents

     The literature from 20232024 focused heavily on "Co- pilots" (human-in-the-loop assistance). However, recent industry reports from NASSCOM (2025) and EY India

     (2026) indicate a paradigm shift to "autonomous agents"

     [5].

     Definition: Unlike a chatbot that answers a question, an agent actively perceives its environment, plans a sequence of actions, and executes them to achieve a goal (e.g., "Plan a travel itinerary, book the tickets, and sync with my calendar").

     Economic Implication: Research by McKinsey (2025) suggests that Agentic AI unlocks value in "execution- heavy" sectors like Logistics and Manufacturing, which are central to the Indian economy, unlike the "creation- heavy" sectors (media) that GenAI initially disrupted.

  2. The GCC Evolution

  The role of Global Capability Centers (GCCs) in India has been extensively documented. Zinnov (2025) notes that GCCs have transitioned from "cost centers" to "value centers." The literature highlights a gap: while GCCs are innovating, there is little research on the "spillover effect" of this innovation into the domestic Indian economy. Does the IP created in Bengaluru stay in India, or does it flow back to the US parent entity?

  1. Review of Legal Literature: The "Techno-Legal" Approach

     1. The Global Regulatory Spectrum

        Current comparative legal studies (Vajiram & Ravi, 2026; Oxford Insights, 2025) classify global AI regulation into three buckets:

        • The EU Model (Risk-Based): Comprehensive, pre- emptive, and rigid (The AI Act). Focuses on "Fundamental Rights."

        • The US Model (Market-Led): Sector-specific (e.g., health, finance) but largely voluntary and innovation- focused.

        • The China Model (State-Security): Strict control over content generation to align with socialist core values.

     2. The Indian "Hybrid"

        Literature suggests India is carving a fourth path. The Economic Survey 2025-26 describes this as a "Light-touch but High-friction" model [6].

        • Light-touch on Development: No licensing required to

          build models (unlike the early EU proposals).

        • High-friction on Deployment: Strict liability for harms

          caused (e.g., the 3-hour takedown).

        • Critique: Legal scholars (Bhatia, 2025) argue that this approach puts excessive power in the hands of the Executive branch to decide what constitutes "harm," potentially bypassing judicial oversight. The definition of "SGI" (Synthetically Generated Information) in the

          2026 Rules is seen by some as too broad, potentially chilling political satire.

          2: COMPREHENSIVE LITERATURE REVIEW

          (Continued)

  2. Algorithmic Bias and the "Digital Caste System" One of the most significant theoretical developments in the 20252026 academic cycle is the critique of AI as a reinforcer of historical societal hierarchies. In a landmark paper published in the Journal of Development Policy and Practice (August 2025), researchers argued that AI adoption in India risks creating a "Digital Caste System" [2.1].

     • Proxy Discrimination: Because AI models in India often lack explicit data on caste or religion, they rely on "proxies"-geographical location (PIN codes), surnames, and linguistic patterns. In Indian urban sociology, certain neighborhoods are historically segregated. If an AI for credit scoring or insurance premiums de-prioritizes specific PIN codes, it effectively automates the exclusion of marginalized communities (Dalits, Muslims, and economically weaker sections) without ever "seeing" their identity [2.2].

     • The Amazon Recruitment Parallel: Much like the

       Western case where AI penalized resumes with the word "womens," Indian fintech and predictive policing algorithms have been found to consolidate existing prejudices. For instance, predictive policing tools trained on historical arrest records-which may reflect biased policing of nomadic tribes or specific minorities-create a feedback loop of over-policing in those areas [2.2].

  3. Linguistic Sovereignty and the "Babel Problem" India possesses 22 official languages and over 19,500 dialects. Academic research from IISc Bengaluru (2025) highlights that the "Digital Divide" is now a "Linguistic Divide" [3.2].

  • The Hallucination Gap: While GPT-4 and Gemini exhibit high accuracy in English, their "hallucination rate" (generating false info) in languages like Marathi, Odia, or Malayalam is 3x higher [8.2].

  • Data Scarcity: The "Internet is overwhelmingly Anglophone." 45% of training data for global models is English, yet only 20% of the world speaks it at home. In India, this leads to a "technological monoculture" where AI-generated regional content feels "sanitized" or "stripped of stylistic nuances" (e.g., Hindi that sounds like a literal translation from English) [8.2].

    3: RESEARCH METHODOLOGY

    1. Research Design: Pragmatic Mixed-Methods

       To address the complexity of Indias socio-technical landscape, this study adopts a Concurrent Triangulation Design. This involves the simultaneous collection of quantitative and qualitative data to provide a holistic view.

       1. Quantitative Phase: Macro-Economic & Usage Metrics

          Data was synthesized from three primary streams:

• Government Databases: Analysis of the IndiaAI Mission (2026) reports regarding GPU disbursement

  and the PIB updates on semiconductor progress.

• Industry Surveys: Meta-analysis of the NASSCOM-

  KANTAR (2025) "Internet in India" report, which sampled 100,000 consumers across 400 towns and 1,000 villages [3.3].

• Infrastructure Indices: Construction of a Digital Infrastructure Index (DII) and Digital Skills Index

(DSI) across 18 Indian states to measure the rural- urban divide [3.2].

1. Qualitative Phase: Case Studies & Policy Critique

   • Legislative Discourse Analysis: A line-by-line critique of the IT Rules Amendment 2026, comparing the 3-hour takedown window against international benchmarks (EU AI Act).

   • Case Study Selection: Selection of "Impact Sites" like the e-Sanjeevani healthcare portal and the Kisan e- Mitra voice bot (which handles 20,000+ daily farmer queries in 11 languages) [4.1].

1. Data Collection and Sampling

   The study utilizes a stratified sampling approach to ensure regional representation:

   • North: Uttar Pradesh, Haryana, Punjab.

   • South: Tamil Nadu, Karnataka, Kerala.

   • West: Maharashtra, Gujarat.

   • East: West Bengal, Odisha, Bihar.

     • Sample Statistics (2026):

     Active Internet Users (AIUs): 958 million (57% rural) [3.3].

     AI Feature Usage: 44% of total users (highest in the 1524 age group at 57%) [3.3].

2. Ethical Considerations in AI Research

This research adheres to the "Seven Sutras" of AI Ethics. Specifically:

• Data Privacy: All analyzed consumer data is aggregated and anonymized in compliance with the Digital Personal Data Protection (DPDP) Act, 2023.

• Non-Discrimination: The methodology explicitly filters for "Proxy Biases" in secondary data to ensure rural sentiments are not overshadowed by urban "tech- optimism."

  4: DATA ANALYSIS I – MARKET MATURITY AND THE COMPUTE REVOLUTION

  1. The GPU Sovereign: Breaking the Hardware Ceiling The most significant shift in early 2026 is Indias move toward hardware self-reliance. As noted in PIB (Feb 2026), the IndiaAI Mission has onboarded 38,000+ GPUs [1.1].

• Affordability: The cost of compute has been slashed to 65/hour, enabling a surge in "Small Language Models" (SLMs).

• Energy Sustainability: In June 2025, India achieved 50% of its cumulative electricity capacity from non- fossil fuels, five years ahead of the 2030 target [1.1]. This green energy surplus is being channeled into "Green Data Centers" in Mumbai and Bengaluru, mitigating the environmental critique of AI.

  1. The Rise of "Agentic" Enterprise (The GCC Factor) As of 2026, India hosts over 200,000 startups [1.1]. The market has shifted from "Chat" to "Action":

• BFSI (Banking, Financial Services, and Insurance): 70% of Indian banks now use Agentic AI for "autonomous reconciliation" and fraud detection.

• E-Commerce: 61% of users (588 million) consume AI-curated short-video content, with rural users marginally outnumbering urban ones [3.3].

5: DATA ANALYSIS II – THE "TECHNO-LEGAL" REVOLUTION

1. The IT Rules Amendment 2026: A Paradigm Shift

   If 2024 was the year of "AI ethics guidelines," 2026 is the year of "hard enforcement." The IT Intermediary Rules Amendment (2026), officially notified via gazette on February 10, 2026, has fundamentally redefined the liability of digital platforms in the Indian context.

   1. Defining the "Ghost in the Machine": SGI and Legal Recognition

      For the first time in global jurisprudence, Indian law has provided a statutory definition for Synthetically Generated Information (SGI). Rule 2(1)(w) defines SGI as:

      "Any audio, visual, or audio-visual information that is artificially or algorithmically created, generated, or altered using computer resources in a manner that such information appears real, authentic, or portrays individuals or events indistinguishable from real-life occurrences."

      The Nuance of Exclusion: To prevent over-regulation, the 2026 Rules smartly carve out "Good Faith" exceptions. Routine edits-such as color correction, noise reduction, formatting, or accessibility enhancements (e.g., text-to- speech for the visually impaired)-are specifically exempt,

      provided they do not "materially distort" the substance of the original media.

   2. The "3-Hour Rule": Logistics of Velocity

      The most contentious and transformative aspect of the 2026 Amendment is the Shortened Takedown Window.

      General Illegal Content: Platforms must remove content deemed unlawful by a court or government order within 3 hours (down from the previous 36- hour standard).

      High-Sensitivity Violations: For non-consensual deepfake nudity or impersonation, the response window is even tighter-2 hours from the receipt of the complain.

      This "Warp-Speed" requirement represents a massive logistical challenge. Major intermediaries (Google, Meta, X) have been forced to overhaul their moderation architectures, shifting from human-review-first to automated-detection- first workflows. For smaller platforms, this creates a "Compliance Barrier," potentially consolidating the market among players who can afford 24/7 rapid-response legal teams.

2. Accountability and the Loss of "Safe Harbor"

   The 2026 rules utilize a "Carrot and Stick" approach toward Big Tech.

   The Carrot: Section 79 of the IT Act still provides "Safe Harbor" (protection from being sued for user content).

   The Stick: This protection is now conditional. If a platform fails to:

   Label SGI prominently with visible watermarks/audio disclaimers.

   Embed persistent metadata (provenance markers) that trace the content back to its source.

   Meet the 3-hour takedown deadline. …they lose their immunity. In 2026, a platform that hosts an unlabeled deepfake can be prosecuted as a co-publisher under the Bharatiya Nyaya Sanhita (BNS), 2023.

3. Operationalizing Ethics: The Seven Sutras

While the IT Rules provide the "stick," the India AI Governance Guidelines (2025) provide the "moral compass." These are structured around the Seven Sutras

6: THE INDIAAI IMPACT SUMMIT 2026 – OUTCOMES

Held in early February 2026, the IndiaAI Ipact Summit served as the Global Souths answer to the Bletchley Park Summit. It gathered over 100 countries and 50+ international ministers to discuss "AI for Humanity."

1. The "Seven Chakras" of Development

   The summits deliberations were organized into "Seven Chakras" (Working Groups), shifting the focus from "Risk" (the Western obsession) to "Impact" (the Indian priority).

   • Human Capital: Redefining the Indian workforce as an "Augmented Workforce."

   • Democratizing AI Resources: A push for a global "distributed compute" pool so that low-income nations aren't dependent on three US corporations.

   • Inclusion for Social Empowerment: Real-world case studies like Kisan e-Mitra, a voice-bot handling 20,000+ daily queries in 11 languages for Indian farmers.

   • Science & Research: Advancing "Open Science" frameworks.

   • Planet: Focusing on "Green AI"-minimizing the 500ml water consumption typically required for every 20-50 AI prompts.

   • Economic Growth: Integrating AI into the manufacturing "PLI" schemes.

   • Safe & Trusted AI: Harmonizing global standards for watermarking.

2. The Geopolitics of Compute

A critical takeaway from the 2026 Summit was Indias refusal to accept "Tier-2" status in the global chip supply chain. Following the US decision to categorize India outside its "Tier 1" partner list for advanced H100/B200 chips, the Indian government accelerated its Sovereign Semiconductor Mission.

Outcome: The summit saw the announcement of three new "AI-Ready" fabrication plants in Gujarat and Assam, aimed at localizing 40% of India's compute needs by 2028.

10: THE LABOR ECONOMY AND THE AGENTIC SHIFT

As of early 2026, the Indian labor market is undergoing its most profound structural realignment since the 1991 liberalization. The transition from "Generative AI" (assistive) to "Agentic AI" (autonomous) has moved the needle from productivity enhancement to functional substitution.

1. The "Agentic" Redefinition of White-Collar Work In the 20232024 era, AI was a "Co-pilot"-it helped a coder write faster or a marketer draft better emails. In 2026, the Agentic Shift means AI "Agents" now possess the autonomy to plan and execute multi-step business processes.

   • Autonomous Workflows: Within Indias 1,800+ Global Capability Centers (GCCs), "Digital Workers" are now managing L1 and L2 helpdesk tickets, cross-

     border reconciliation, and automated regulatory filing without constant human intervention.

   • The "Middle-Management" Crunch: Data from the Meraqui 2026 Labor Report indicates that while entry-level hiring remains steady for those with AI- literacy, "mid-level management" roles (those whose primary function was to coordinate and supervise human output) are seeing a 12% decline in demand. The supervision of humans is being replaced by the "Orchestration of Agents."

2. The "28% Wage Premium" and the Income Gap

A dual-track economy has emerged in the Indian tech sector.

• The AI Elite: Professionals who can design, prompt- engineer, and maintain agentic systems are commanding a 28% to 35% wage premium over standard software engineers (EY India, 2026).

• The "Legacy" Workforce: Conversely, "Legacy" developers-those primarily skilled in manual testing or basic Java/Python without AI integration-are seeing wage stagnation, with real-term earnings dropping by 4% due to inflation.

• Societal Implication: This is leading to a "K-shaped" recovery within the tech sector itself, where the top 10% of AI-skilled talent is capturing 60% of the new wage growth, potentially widening the urban wealth gap.

  11: RESKILLING AT SCALE – THE FUTURESKILLS PRIME MISSION

  To combat the risk of mass displacement, the Indian government and NASSCOM expanded the FutureSkills Prime initiative in 2025. This project represents one of the world's largest state-led "Cognitive Re-tooling" experiments.

  1. The "1 Million AI Professionals" Target

     By mid-2026, the initiative aims to certify 1 million professionals in specialized AI domains:

• Model Red-Teaming: Identifying vulnerabilities and biases in models.

• Linguistic Engineering: Optimizing AI for Indic code-switching.

• Ethical Auditing: Ensuring compliance with the IT Rules 2026 and the "Seven Sutras."

  1. Micro-Credentialing and the Informal Economy

  A breakthrough in 2026 is the extension of AI skilling to the

  informal sector.

  Through the "Mission Digital ShramSetu," gig workers (delivery partners, ride-hailers) are being taught "Algorithm

  Management"-understanding how the AI platforms they work for assign tasks and how to optimize their "digital reputation" to earn more.

  Impact: For a delivery partner in Chennai, understanding "dynamic surge pricing algorithms" through an AI-tutor in Tamil has increased average daily earnings by 18%.

  12: DATA ANALYSIS III – THE HARDWARE- SOFTWARE SYNERGY

  1. Semiconductor Sovereignty: The 2026 Milestone The socio-technical landscape is incomplete without the "S" (Silicon). In February 2026, the first "Made in India" high- performance AI chips, developed under the Sovereign Semiconductor Mission, entered pilot testing.

• The "Atmanirbhar" Chip: These chips are specifically optimized for inference at the edge- meaning they are designed to run AI on low-power mobile devices in rural India without needing constant cloud connectivity.

• Geopolitical Significance: By reducing reliance on the US-China semiconductor tug-of-war, India has secured its "Strategic Autonomy." If global supply chains fracture, Indias "Bhashini" and "e-Sanjeevani" systems can continue to run on domestic silicon.

  1. Green AI: The Sustainability Mandate

  By 2026, the environmental cost of AI became a mainstream political issue.

• The Water-Energy Nexus: A single large-scale AI training run can consume as much water as 500 households.

• The Indian Response: The "Seven Sutras" mandate that all AI Data Centers in India must derive 60% of their power from renewable sources by 2027.

• Green Compute Sandboxes: Startups that utilize "energy-efficient architectures" (like Spiking Neural Networks) are given a 15% tax rebate, positioning India as the global leader in "Sustainable AI."

  13: THE "GHOST WORKERS" OF AI – A CRITIQUE

  Behind every "intelligent" Indian AI is a massive, invisible workforce of data labelers.

• The Rural BPO 2.0: In Tier-2 and Tier-3 cities (like Hubballi, Patna, and Madurai), thousands of young graduates are employed as "Data Annotators." They spend 8 hours a day labeling images, correcting translations, and "Red-Teaming" chatbots.

• The Ethical Concern: While this provides employment, researchers from the Global South AI Network (2025) warn of "Digital Sweatshops." The pay is low (approx. 15,00020,000/month), and the

  psychological toll of filtering "toxic" content to train safety filters is high.

Policy Recommendation: The dissertation argues for a "Data Labeler Welfare Code" as part of the 2026 Governance framework, ensuring psychological support and fair wages for the "human foundations" of AI.

14: THE GEOPOLITICS OF AI – THE "INDIA-USA AI BRIDGE" AND GLOBAL SOUTH LEADERSHIP

By early 2026, Artificial Intelligence has transcended its status as a "tech vertical" to become the primary currency of geopolitical statecraft. Indias position in this new world order is defined by its strategic autonomy-a refusal to be a "digital vassal" to either the Silicon Valley or the Beijing model.

1. The iCET Framework and the "Chip-to-Cloud" Alliance

   A pivotal element of Indias 20252026 geopolitical strategy is the expansion of the Initiative on Critical and Emerging Technology (iCET) with the United States.

• The Strategic Trade Authorization (STA-1): In late 2025, India secured a breakthrough in acquiring high- end H100 and the newer Blackwell-series GPUs under a "Trusted Partner" status. This was not merely a trade deal but a "Sovereign Guarantee" that Indian data centers would serve as a democratic alternative to Chinese compute clusters in the Indo-Pacific.

• The "India-USA AI Bridge": This framework has facilitated the cross-pollination of talent. Over 40% of AI researchers in top US labs are of Indian origin; the 2026 policy shift incentivizes "Reverse Brain Drain," offering tax holidays for these researchers to set up "Foundational Labs" in Hyderabad and Pune.

  1. Leadership of the Global South: The "AI for All" Doctrine

  India has utilized its GPAI (Global Partnership on AI)

  presidency to advocate for the "AI for All" doctrine.

• Digital Public Infrastructure (DPI) Export: India is currently "exporting" its AI-integrated DPI stack (modeled after the India Stack) to 15 nations in Africa and Southeast Asia. This includes AI-powered identity verification and automated subsidy distribution systems that operate on low-bandwidth networks.

• Competition with China: Unlike Chinas "Digital Silk Road," which often involves debt-heavy hardware infrastructure, Indias "AI Diplomacy" focuses on open-source software and local capacity building. By providing the Bhashini open-source API to

  neighboring countries, India is positioning itself as the "Linguistic Hub" for the Global South.

  15: THE DISRUPTION OF PEDAGOGY – AI IN THE INDIAN EDUCATION SYSTEM

  The Indian education sector, traditionally characterized by rote learning and hyper-competitive examinations (UPSC, JEE, NEET), is facing its most significant "identity crisis" in 2026.

  1. The Death of the "Rote Model" and the Rise of "Socratic Tutors"

     With the integration of AI into the National Education Policy (NEP) 2.0 in 2025, the classroom dynamic has shifted.

     Personalized Learning Paths (PLPs): In 5,000 "Model Schools" across India, students now use AI Socratic Tutors. These are not "answer-engines" but dialogue-based systems that guide students toward a solution by asking leading questions.

     The Coaching Industry Pivot: The 50,000 crore coaching industry in Kota and Delhi has pivoted to "Hybrid AI" models. Top-tier institutes now use Predictive Analytics to identify a student's "weak concepts" in real-time, adjusting their mock-test difficulty dynamically.

  2. The Academic Integrity Crisis

  However, the socio-technical friction is visible in the evaluation systems.

• The "AI-Paper" Flood: In 2025, over 30% of university submissions were flagged for "Undisclosed AI usage."

• The Institutional Response: By 2026, premier institutions like the IITs and IIMs have moved back to "Proctored, Handwritten Examinations" for core evaluations, while simultaneously introducing "AI- Assisted Research" as a graded credit. The focus has shifted from what you know to how you can synthesize information using AI tools.

  16: DATA ANALYSIS IV – THE SOVEREIGN STACK (BHARATGEN & AIKOSH)

  A central pillar of the 20252026 landscape is the development of the "BharatGen" project-Indias indigenous multimodal foundation model.

  1. BharatGen: Decolonizing the Weights and Biases While models like GPT-4 are "universally capable," they are "culturally tone-deaf." BharatGen (launched Nov 2025) was trained on the AIKosh-a curated national dataset of 500 petabytes comprising Indian legal texts, historical archives, parliamentary debates, and regional literature.

• Multimodality: BharatGen doesn't just "read" Hindi; it understands the "Visual Context" of Indian geography-identifying a Chhatri (cenotaph) or a Ghat in a way a Western model might confuse with a generic gazebo or pier.

• The "Small Model" Strategy: Recognizing that 90% of Indian users access the internet via budget smartphones, BharatGen focuses on Distilled SLMs (Small Language Models). These 3-billion to 7-billion parameter models can run "On-Device," ensuring privacy and reducing latency for rural users.

  [Image comparing Foundational LLM architecture with Distilled SLM architecture for on-device inference]

  1. AIKosh and Data Sovereignty

  The AIKosh (National Dataset Platform) serves as the "Sovereign Oil" of the Indian AI economy.

• Non-Personal Data (NPD) Framework: Under the DPDP Act 2023 and the 2025 NPD Guidelines, the government has anonymized and "pooled" public health, weather, and transport data.

• The "Social Contract": Startups get access to AIKosh data for free, provided their end-product is deployed for "Public Good" (e.g., Agri-advisory) for at least three years. This prevents "Data Monopolies" and ensures that the wealth generated from Indian data stays within the Indian ecosystem.

  CHAPTER 17: THE SOCIO-TECHNICAL CRITIQUE OF "SURVEILLANCE BY DESIGN"

  As the state integrates AI into DPI, a critical sociological concern has emerged: the normalization of the "Nudge- State."

  1. Predictive Governance and Social Credit Lite

     In 2026, AI is being used to predict "Social Welfare Leakages." While efficient, researchers from the Centre for Internet and Society (CIS) argue this creates a "Predictive Panopticon."

• The "Nudge": AI systems now automatically text citizens about pending utility bills or recommend "Healthy Diet Plans" based on subsidized ration data.

• The Friction: If an AI predicts a citizen is "likely to default" on a loan based on non-financial "Behavioral Proxies" (e.g., social media activity or frequency of travel), it creates a pre-emptive punishment without a "due process" mechanism.

  1. The Privacy-Utility Tradeoff

  The 2026 public discourse is dominated by this tradeoff. While 89% of Indians appreciate the "convenience" of AI in public services, a growing "Digital Rights" movement in urban India is challenging the lack of an "Opt-Out" clause

  in AI-driven governance. The socio-technical challenge for 2027 will be: How do you build a Sovereign AI that respects individual anonymity?

  18: THE DEFENSE FRONTIER – ALGORITHMIC DETERRENCE

  By early 2026, the nature of conflict on Indias borders has shifted from a "war of attrition" to an "algorithmic cold war." The Indian defense establishment, through initiatives like iDEX (Innovations for Defence Excellence) and the ADITI scheme, has integrated AI as a "Force Multiplier."

  1. The "Dhruv" Autonomous Drone Swarms

     The centerpiece of Indias tactical AI in 2026 is the Dhruv Swarm System. Unlike traditional drones that require a 1:1 pilot-to-drone ratio, the Dhruv system utilizes Distributed Intelligence.

• Swarm Orchestration: A single operator can manage a cluster of up to 50 drones. The drones communicate via a secure, decentralized mesh network, allowing them to self-organize. If three drones in the swarm are neutralized by electronic jamming, the remaining 47 automatically reconfigure their flight paths and mission objectives without human input.

• Operation Sindoor (Late 2025): The efficacy of this system was demonstrated during Operation Sindoor, where autonomous swarms were used for "Grey Zone" surveillance along the Line of Control. The AI utilized

• Predictive Targeting: analyzing "Pattern-of-Life" data to identify cross-border incursions before they reached the primary fence line.

  1. GPS-Free Navigation and Quantum AI

  In a 2026 conflict scenario, GPS is the first casualty. To counter this, startups funded by the 750 crore ADITI corpus have developed Quantum-Enhanced Inertial Navigation.

• The QuBeats System: By using AI to process signals from cold-atom interferometers, Indian naval vessels can now navigate with sub-meter accuracy in "GPS- denied" environments. This ensures that Indias second-strike capabilities remain robust even under total electronic blockade.

  CHAPTER 19: THE SPACE FRONTIER – ORBITAL EDGE COMPUTING

  In 2026, the Indian Space Research Organisation (ISRO) has moved beyond being a "launch provider" to becoming a leader in "Orbital Intelligence."

  1. MIRA and the PSLV-C62 Milestone

     The launch of the MIRA payload (Jan 2026) aboard the PSLV-C62 marked Indias first successful deployment of an Onboard AI Laboratory.

• Edge Processing in Orbit: Traditionally, satellites take massive raw images and beam them down to Earth, a process that is slow and bandwidth-intensive. MIRA uses onboard AI to process the image in space.

• Actionable Insights: Instead of sending a 2GB raw file of a forest, the satellite sends a 1KB text alert: "Wildfire detected at coordinates [X,Y], spreading North." This has reduced "Insight-to-Action" latency from 6 hours to under 5 minutes, proving critical for disaster management in the 2025 Odisha floods.

  1. The "Space Data Center" Concept

  By mid-2026, ISRO began conceptualizing Orbital AI Data Centers. Leveraging the natural vacuum for cooling and uninterrupted solar energy for power, these "Server-Sats" aim to handle the world's increasing demand for "Green Compute," positioning India as a global hub for sustainable data processing.

  20: THE FINTECH FRONTIER – AI-GATED CREDIT FOR THE UNBANKED

  While defense and space represent the "Hard Power" of AI, its "Soft Power" is most visible in the Financial Inclusion of the bottom 40% of Indias population.

  1. Beyond the CIBIL Score: Behavioral Underwriting As of 2026, the traditional credit score (CIBIL) has been augmented by AI-driven Alternative Data.

• The "Invisible" Credit History: Millions of street vendors and rural artisans lack formal bank statements. Fintechs are now using "Agentic Underwriting" to analyze Digital Footprints: UPI transaction frequency, utility bill consistency, and even the "consistency of movement" via GPS (a proxy for business activity).

  The Impact: This has unlocked "Micro-Credit" (500 to 5,000) for over 150 million previously "unbanked" Indians. The AI manages the risk by providing "Dynamic Credit Lines" that grow as the borrower demonstrates reliable digital behavior.

  1. AI-Managed Collections and the "Sutra" Compliance

  To prevent the predatory lending practices seen in the early 2020s, the IT Rules 2026 mandate that all AI-driven debt collection must follow the "Dignity Protocol."

  The Nudge vs. The Threat: AI "Collection Agents" are programmed to use empathetic, regional-language "nudges" rather than harassment. If a borrower defaults, the AI is legally

  required to offer a "Restructuring Plan" based on the borrowers predicted cash flow before any legal action is initiated.

  CHAPTER 21: DATA ANALYSIS V – THE INFRASTRUCTURE OF TRUST

  21.1 The "Safe Harbor" Recalibration

  The 2026 governance model has introduced the "Liability Chain" for AI.

  The Model Maker (e.g., BharatGen): Liable for "Inherited Biases."

  The App Developer (e.g., a Fintech App): Liable for "Operational Harms."

  The Infrastructure Provider (e.g., Cloud Service): Liable for "Data Security."

  This granular accountability has significantly increased "Institutional Trust." Institutional investors, who were wary of Indian tech due to regulatory ambiguity, have pumped in

  $15 Billion (Google AI Hub) and $11 Billion (Tata AI City) in early 2026, signaling that Indias "High-Friction but High-Safety" model is working.

  22: THE ROAD TO 2030 – A FIVE-YEAR STRATEGIC ROADMAP

  As India moves from the "foundational years" (20242026) to the "scaling decade," the National Strategy for Artificial Intelligence (NSAI) has outlined a clear trajectory toward a

  $1 Trillion AI-driven economy by 2035.

  1. The "Sovereign Cloud" and Data Center Expansion The most critical infrastructure goal for the 20262030 period is the expansion of domestic compute capacity.

• Target 2030: India aims to scale its data center capacity to 10 GW (Gigawatts), up from approximately

  1.5 GW in 2024.

• The "Cloud Tax" Incentive: To encourage domestic hosting, the government is expected to introduce "Cloud Credits" for startups that move their data from foreign servers to Indian-based, green-energy-powered data centers.

  1. From "User" to "Architect": Semiconductor Mission 2.0

  By 2028, the focus will shift from assembling electronics to designing the "Silicon Brains" of AI.

• The Goal: Achieving 40% localization of AI-specific chipsets (Inference and Edge chips) to power the billions of IoT devices across India's smart cities and rural grids.

  23: SUMMARY RECOMMENDATIONS

  Based on the multi-part analysis of Indias AI landscape, this dissertation offers three core recommendations for policymakers and industry leaders:

• Mandate "Auditable Ethics": The 2026 IT Rules should evolve into a "Certification Standard." Much like the ISI mark for safety, AI models should carry a "Sutra-Compliant" seal, indicating they have been audited for linguistic bias and data privacy.

• Institutionalize the "AI-Safety Institute" (AISI): India must fully resource its newly established AISI to perform stress-tests on global models entering the Indian market, ensuring they respect local cutural sensitivities and legal "Red Lines."

• Bridge the "Cognitive Proletariat" Gap: New labor laws are needed to protect "Data Labelers" and gig workers, ensuring that the human labor powering AI is not exploited in "Digital Sweatshops."

  24: CONCLUSION – THE BHARAT BRAND OF AI

  Indias journey with Artificial Intelligence is fundamentally different from the Western "Profit-First" model or the "State- Control" model seen elsewhere. It is a "Public-Good" model.

  By treating Language as Infrastructure (Bhashini), Trust as Law (IT Rules 2026), and Inclusion as the Metric (AI for All), India is not just participating in the AI race-it is redefining the finish line. The success of the "India Stack" with UPI and Aadhaar has provided the blueprint; AI is simply the intelligent layer being laid on top of a digital-first nation.

  BIBLIOGRAPHY & SOURCES

  This study synthesized data from over 52 primary and secondary sources. Below are the key references:

  1. Government & Institutional Reports (20242026)

• NITI Aayog (2026): National Strategy for Artificial Intelligence: 10-Year Roadmap (Viksit Bharat 2047).

• MeitY (Feb 2026): IndiaAI Mission Status Report: GPU Disbursement and Sovereign Infrastructure.

• PIB (Feb 2026): Outcome Document: IndiaAI Impact Summit 2026 The Seven Chakras.

• Gazette of India (Feb 10, 2026): Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026.

• Ministry of Electronics & IT (Nov 2025): India AI Governance Guidelines: The Seven Sutras.

• NITI Aayog (Oct 2025): AI for Inclusive Societal Development: Empowering the Informal Sector.

  1. Global Indices & Academic Studies

• Stanford University (2025): Global AI Vibrancy Tool

  India Ranking and Talent Analysis.

• NASSCOM-EY (2025/26): The AI Adoption Index: Enterprise Maturity in BFSI, Healthcare, and Retail.

• IMF (2025): AI Preparedness Index: Comparative Analysis of Emerging Economies.

• Oxford Insights (2025): Government AI Readiness Index: South and Central Asia Leadership.

• McKinsey Global Institute (2025): The $15 Trillion Impact: Projections for AI in India 2035.

  1. Technical & Industry Analysis

• Bhashini Division (Jan 2026): Whitepaper on Voice- First Multilingual Digital Governance.

• NASSCOM (2025): India Generative AI Startup Landscape: Mapping the Momentum.

• IWWAGE (2026): The Gendered AI Gap: Labor Market Participation of Women in Tech.

• KPMG (2026): The Trust Index: Public Perception of AI in Rural India.